Security Think Tank: In the cloud, anti-human approaches set us up to fail